This table displays the selected game’s current values on four BGG outcomes (UsersRated, Average, GeekRating, Weight) along with my predictive model(s) estimated values for where these games are likely to end up.
Published | ID | Name | Type | UserRatings | Average | GeekRating | Weight |
2021 | The Third Winter: The Battle for the Ukraine September 1943-April 1944 | Current | 36 | 8.97 | 5.55 | 3.83 | |
Estimated | 200 | 8.64 | 5.81 | 4.49 |
To see more information about the game on boardgamegeek, click on the game’s ID or Name to go straight to the game’s profile page.
Description |
The Third Winter: The Campaign for the Ukraine, September 1943 - April 1944, begins with the “Race to the Dnepr,” as Army Group South rushed to a secure defensive position behind one of Europe’s largest rivers. Four powerful Soviet Fronts must be used in close cooperation to overcome the Axis forces, which include two Panzer Armies. The game contains 85% of the total Wehrmacht’s mechanized strength at this stage of the war and 75% of the Red Army’s. The Germans must hold the Dnepr barrier as long as possible, followed by a fighting retreat to the Carpathian Mountains to the west. The Soviets can crush the Axis invader with the advantage of their most reliable ally: General Winter. It is the Third Winter on the Russian Front. |
Information on the game’s designers, artists, families, mechanics and categories from BGG.
Published | ID | Name | Publisher(s) | Designer(s) | Artist(s) | Categories | Families | Mechanics |
2021 | The Third Winter: The Battle for the Ukraine September 1943-April 1944 | Multi-Man Publishing | Antony Birkett | Curtis Baer | Wargame | Country: Germany | Dice Rolling | |
Dean Essig | Nicolás Eskubi | World War II | Country: Soviet Union | Grid Movement | ||||
Chip Saltsman | Dean Essig | Country: Ukraine | Hexagon Grid | |||||
History: World War II (Eastern Front) | Simulation | |||||||
Series: Operational Combat Series (MMP) | Zone of Control | |||||||
For the full profile of the selected game, click on the links in the table or go to https://boardgamegeek.com/boardgame/288572
For each of the predictions above, I used models that were trained on historical boardgamegeek data in order to predict the selected game. How did the model(s) arrive at their predictions? The following plot displays Shapley values to indicate what features were most influential for the model’s predictions. Anything in blue increased the model’s predictions, anything in red decreased the model’s predictions.
What games are similar to The Third Winter: The Battle for the Ukraine September 1943-April 1944? I compare the selected game to every other game on boardgamegeek based on its mechanics, playing time, complexity, and game type. The tables below display the games which had the most similar data to The Third Winter: The Battle for the Ukraine September 1943-April 1944. The reported similarity is the cosine similarity between the selected game and all other games published before 2021.
This similarity score ranges from -1 to 1, where similarity of 1 indicates that games are exactly identical (a game will have a score of 1 with itself).
Note: This analysis does not look at a game’s ratings on BGG, publisher, artist, or designer - it focuses on finding similar games based on how they play, not how they are rated or by their theme.
The table below displays games with the most similar data to The Third Winter: The Battle for the Ukraine September 1943-April 1944. Click on a game’s ID or Name to go straight to its profile on boardgamegeek.
Rank | Published | ID | Name | Similarity | GeekRating | Average | Weight | UserRatings |
-- | 2021 | The Third Winter: The Battle for the Ukraine September 1943-April 1944 | 1.00 | 5.55 | 8.97 | 4.49 | 36 | |
1 | 2019 | 0.99 | 5.51 | 6.58 | 4.25 | 36 | ||
2 | 2011 | 0.99 | 5.58 | 7.72 | 3.61 | 80 | ||
3 | 2004 | The Cossacks Are Coming!: The Tannenberg Campaign August and September, 1914 – 2nd Edition | 0.99 | 5.53 | 7.25 | 4.31 | 34 | |
4 | 2016 | 0.98 | 5.53 | 7.04 | 4.33 | 45 | ||
5 | 1984 | 0.98 | 5.76 | 7.50 | 4.41 | 306 | ||
6 | 1994 | 0.98 | 5.76 | 7.63 | 4.40 | 269 | ||
7 | 2007 | 0.97 | 5.53 | 6.60 | 2.86 | 69 | ||
8 | 2016 | 0.97 | 5.65 | 8.45 | 3.08 | 119 | ||
9 | 2018 | 0.97 | 5.62 | 9.00 | 4.20 | 105 | ||
10 | 2015 | 0.97 | 5.75 | 8.21 | 4.16 | 210 | ||
11 | 1979 | 0.97 | 5.58 | 7.08 | 3.18 | 107 | ||
12 | 1995 | 0.97 | 5.59 | 6.81 | 3.67 | 146 | ||
13 | 2004 | 0.97 | 5.58 | 6.56 | 3.68 | 153 | ||
14 | 1994 | 0.97 | 5.49 | 5.49 | 2.92 | 94 | ||
15 | 2020 | 0.96 | 5.59 | 7.68 | 2.50 | 97 | ||
16 | 2010 | 0.96 | 5.51 | 6.23 | 3.25 | 33 | ||
17 | 1991 | Civil War Classics, Volume 1: The Battles of Pea Ridge & Shiloh | 0.96 | 5.46 | 3.25 | 3.12 | 34 | |
18 | 1985 | 0.96 | 5.52 | 6.25 | 2.62 | 67 | ||
19 | 1997 | Eastwall: Battles for the Dnepr, September 1943 - February 1944 | 0.96 | 5.51 | 5.94 | 3.00 | 36 | |
20 | 1986 | 0.96 | 5.49 | 5.40 | 3.91 | 79 | ||
21 | 1983 | 0.96 | 5.58 | 7.09 | 3.12 | 101 | ||
22 | 2020 | 0.96 | 5.59 | 8.49 | 4.43 | 68 | ||
23 | 2001 | 0.96 | 5.52 | 6.23 | 2.67 | 54 | ||
24 | 2018 | 0.96 | 5.55 | 7.66 | 4.00 | 49 | ||
25 | 2000 | 0.96 | 5.94 | 7.84 | 3.55 | 442 |
Click on the ID or Name for any game to go its profile page on boardgamegeek.
What games are similar to The Third Winter: The Battle for the Ukraine September 1943-April 1944 while also being highly rated by the boardgamegeek community? In this table I multiply every neighboring game’s similarity score by its GeekRating. The resulting score is an attempt to balance these two quantities.
Rank | Published | ID | Name | Score | Similarity | GeekRating | Average | Weight | UserRatings |
-- | 2021 | The Third Winter: The Battle for the Ukraine September 1943-April 1944 | 5.55 | 1.00 | 5.55 | 8.97 | 4.49 | 36 | |
1 | 2011 | 5.87 | 0.90 | 6.55 | 7.79 | 3.09 | 1,571 | ||
2 | 2000 | 5.71 | 0.96 | 5.94 | 7.84 | 3.55 | 442 | ||
3 | 2006 | 5.66 | 0.90 | 6.28 | 7.67 | 2.56 | 1,057 | ||
4 | 1984 | 5.65 | 0.98 | 5.76 | 7.50 | 4.41 | 306 | ||
5 | 1994 | 5.61 | 0.98 | 5.76 | 7.63 | 4.40 | 269 | ||
6 | 1984 | 5.60 | 0.94 | 5.93 | 7.80 | 3.47 | 459 | ||
7 | 2015 | 5.57 | 0.97 | 5.75 | 8.21 | 4.16 | 210 | ||
8 | 2011 | 5.55 | 0.99 | 5.58 | 7.72 | 3.61 | 80 | ||
9 | 1995 | 5.51 | 0.96 | 5.75 | 7.02 | 2.60 | 374 | ||
10 | 2016 | 5.50 | 0.97 | 5.65 | 8.45 | 3.08 | 119 | ||
11 | 2019 | 5.49 | 0.92 | 5.94 | 8.46 | 3.55 | 366 | ||
12 | 2004 | The Cossacks Are Coming!: The Tannenberg Campaign August and September, 1914 – 2nd Edition | 5.49 | 0.99 | 5.53 | 7.25 | 4.31 | 34 | |
13 | 2019 | 5.48 | 0.99 | 5.51 | 6.58 | 4.25 | 36 | ||
14 | 2016 | 5.45 | 0.95 | 5.77 | 8.44 | 4.29 | 200 | ||
15 | 2018 | 5.45 | 0.97 | 5.62 | 9.00 | 4.20 | 105 | ||
16 | 2016 | 5.44 | 0.98 | 5.53 | 7.04 | 4.33 | 45 | ||
17 | 2012 | Red Winter: The Soviet Attack at Tolvajärvi, Finland – 8-12 December 1939 | 5.43 | 0.90 | 6.03 | 7.99 | 2.98 | 492 | |
18 | 2014 | 5.43 | 0.93 | 5.81 | 7.95 | 3.55 | 280 | ||
19 | 1996 | Stalingrad Pocket: The Wehrmacht's Greatest Disaster – 2nd Edition | 5.42 | 0.93 | 5.84 | 7.36 | 2.65 | 437 | |
20 | 2006 | 5.41 | 0.92 | 5.88 | 7.56 | 2.73 | 416 | ||
21 | 1995 | 5.41 | 0.97 | 5.59 | 6.81 | 3.67 | 146 | ||
22 | 1979 | 5.40 | 0.97 | 5.58 | 7.08 | 3.18 | 107 | ||
23 | 1976 | 5.40 | 0.93 | 5.83 | 7.47 | 4.17 | 383 | ||
24 | 2004 | 5.40 | 0.97 | 5.58 | 6.56 | 3.68 | 153 | ||
25 | 2020 | 5.40 | 0.96 | 5.59 | 7.68 | 2.50 | 97 |
I first ran a dimension reduction method (PCA) to identify the main points of variation for games on BGG. The first 25 components explained roughly 50% of the overall variation in the dataset, so I computed the cosine similarity between every game’s score on these 25 components.
The following plot shows every game on BGG plotted on the first two principal components of variation. I then highlight where the selected game and its nearest neighbors fall on these first two components.
Similarity scores are based on 25 principal components - the next table displays the selected game and its neighbors profiles on each of these components.
## Warning in abbreviate(neighbor_name, 60): abbreviate used with non-ASCII chars
What are these components? The following plot shows which features contribute the most to each component.